A faster fast fourier transform

Gilbert Strang, author of the classic textbook Linear Algebra and Its Applications, once referred to the fast Fourier transform, or FFT, as "the most important numerical algorithm in our lifetime." No wonder. The FFT is used to process data throughout today's highly networked, digital world. It allows computers to efficiently calculate the different frequency components in time-varying signals- and also to reconstruct such signals from a set of frequency components. You couldn't log on to a Wi-Fi network or make a call on your cellphone without it. So when some of Strang's MIT colleagues announced in January at the ACM-SIAM Symposium on Discrete Algorithms that they had developed ways of substantially speeding up the calculation of the FFT, lots of people took notice.